RDMA resource leakage detection and reporting

Abstract

Mechanisms are provided for remote direct memory access (RDMA) resource leak detection. A user space context is generated comprising a user space RDMA resource hierarchical data structure. A kernel context is generated comprising a kernel RDMA resource hierarchical data structure. The kernel RDMA resource hierarchical data structure comprises nodes of the users space RDMA resource hierarchical data structure. A request to close a RDMA resource is received and the user space RDMA resource hierarchical data structure is traversed to determine whether the RDMA resource has child RDMA resources allocated in the user space RDMA resource hierarchical data structure. A resource leak is detected in response to determining, based on at least one of the user space RDMA resource hierarchical data structure or the kernel RDMA resource hierarchical data structure, that the RDMA resource has a child RDMA resource allocated in the user space RDMA resource hierarchical data structure.

What is claimed is: 
     
       1. A method, in a data processing system comprising a processor and a memory, for remote direct memory access (RDMA) resource leak detection, the method comprising:
 generating, by the data processing system, in a user space of the data processing system, a user space context comprising a user space RDMA resource hierarchical data structure; 
 generating, by the data processing system, in a kernel space of the data processing system, a kernel context comprising a kernel RDMA resource hierarchical data structure, wherein the kernel RDMA resource hierarchical data structure comprises nodes of the users space RDMA resource hierarchical data structure; 
 receiving, by the data processing system, a request to close a RDMA resource; 
 traversing, by the data processing system, the user space RDMA resource hierarchical data structure to determine whether the RDMA resource has child RDMA resources allocated in the user space RDMA resource hierarchical data structure; 
 detecting, by the data processing system, a resource leak in response to determining, based on at least one of the user space RDMA resource hierarchical data structure or the kernel RDMA resource hierarchical data structure, that the RDMA resource has a child RDMA resource allocated in the user space RDMA resource hierarchical data structure; and 
 reporting, by the data processing system, the resource leak. 
 
     
     
       2. The method of  claim 1 , wherein at least one of the user space RDMA resource hierarchical data structure or the kernel RDMA resource hierarchical data structure comprises a red-black tree data structure. 
     
     
       3. The method of  claim 1 , further comprising:
 enforcing an ordered deallocation policy specifying that parent resources in the user space RDMA resource hierarchical data structure or kernel RDMA resource hierarchical data structure cannot be deallocated prior to child resources of the parent resources being deallocated, wherein detecting a resource leak comprises detecting an attempt to violate the ordered deallocation policy. 
 
     
     
       4. The method of  claim 1 , wherein the kernel RDMA resource hierarchical data structure is a superset of the user space RDMA resource hierarchical data structure and comprises data representing RDMA resources represented in the user space RDMA resource hierarchical data structure and kernel space RDMA resources. 
     
     
       5. The method of  claim 4 , wherein nodes of the kernel RDMA resource hierarchical data structure corresponding to user space RDMA resources or kernel space RDMA resources further comprise associated forensic data collected from the user space RDMA resources or kernel space RDMA resources, wherein the forensic data associated with a node of the kernel RDMA resource hierarchical data structure provides information identifying a source of a resource leak associated with the node. 
     
     
       6. The method of  claim 5 , wherein the forensic data comprises at least one of a process identifier for a process that allocated a RDMA resource associated with the node, an indicator of when the RDMA resource associated with the node was allocated, or an entire call stack of a process that allocated the RDMA resource associated with the node. 
     
     
       7. The method of  claim 5 , wherein reporting the resource leak further comprises:
 correlating the child RMDA resource with a corresponding node in the kernel RDMA resource hierarchical data structure; 
 retrieving from the node corresponding to the child RDMA resource, forensic data; and 
 generating a resource leak notification comprising the forensic data. 
 
     
     
       8. The method of  claim 1 , wherein the user space RDMA hierarchical data structure and kernel RDMA hierarchical data structure are generated using logic that enforces a rule that a queue pair (QP) RDMA resource is a child of a corresponding completion queue (CQ) RDMA resource and a corresponding protection domain (PD) RDMA resource. 
     
     
       9. The method of  claim 1 , wherein the user space RDMA hierarchical data structure and kernel RDMA hierarchical data structure are generated using logic that enforces a rule that all RDMA resources are children of a corresponding protection domain (PD) RDMA resource. 
     
     
       10. The method of  claim 1 , wherein the user space RDMA hierarchical data structure and kernel RDMA hierarchical data structure are generated using logic that enforces a rule that a Fast Registration Page List (FRPL) RDMA resource is a child of a corresponding Fast Registration Memory Region (FRMR) RDMA resource and a RDMA control structure (RCS).